Investigating research strategies for accessing knowledge stored in semantic models

Abstract

Current data storage and retrieval strategies usually use keywords and are not well suited to retrieving spatial configurations, the proportions of rooms or their interrelationships. Instead of using text-based research, a graphical inquiry and query system is proposed that can recognise formal structures on the one hand and concept sketches on the other. Using artificial intelligence methods and multimodal interaction, knowledge is stored in semantic models. From previously stored planning solutions in a BIM, semantic fingerprints are derived that describe their functional and topological characteristics. The search system likewise derives a semantic fingerprint from the spatial configuration of a concept sketch and compares it with fingerprints stored in the repository. Similar matches are then shown to the designer.